Antiderivatives Of Trig Functions

Getting the Basics Right

The standard antiderivatives you need to memorize are short, and I mean short. The integral of sin(x) dx is -cos(x) + C. The integral of cos(x) dx is sin(x) + C. The integral of sec^2(x) dx is tan(x) + C. The integral of csc^2(x) dx is -cot(x) + C. Those four cover more ground than most students realize in an introductory course. The other two, the integral of sec(x)tan(x) dx = sec(x) + C and the integral of csc(x)cot(x) dx = -csc(x) + C, show up less often but are still fair game on exams. What trips people up isn't the basic list itself. It's the sign errors. You integrate cosine and get sine, which feels fine. You integrate sine and get negative cosine, and that negative sign disappears into the background without you noticing it. I've graded enough student work to know this is the single most common mistake, and it costs them points on problems that are otherwise trivial.

Antiderivatives Of Trig Functions In Practice

Here's the part that doesn't get enough attention: these formulas work in either direction. When you're doing antidifferentiation, you're looking for a function whose derivative gives you your integrand. When you're differentiating, you already know the answer. The relationship is bidirectional, and recognizing that early saves you from treating each problem as brand-new material instead of recognizing patterns. One thing I ran into recently that I wish someone had warned me about involves integrands where the trig functions are disguised inside other expressions. I was working through a problem that looked like it needed integration by parts — something involving x times a trig function — and I kept hitting dead ends. After about twenty minutes of wrestling with it, I realized the integrand could be rewritten using a double angle identity first, which collapsed the whole thing into something that only required the basic antiderivative formulas. That kind of simplification step is easy to miss when you're focused on mechanical methods rather than examining what the integrand actually is. Another edge case that catches people off guard involves the integral of secant. The standard antiderivative is ln|sec(x) + tan(x)| + C, but the derivation isn't obvious from any of the basic formulas. The trick is multiplying by (sec(x) + tan(x))/(sec(x) + tan(x)), which is just 1 written in a convenient form. When you see a secant-only integrand, that multiplication by the conjugate-style expression is the move to make. Students who don't know this trick either leave the integral unsolved or try to force a substitution that doesn't work.

The power-reduction and product-to-sum formulas are where things get genuinely useful. If you encounter an integral like sin^2(x) dx or cos^2(x) dx, the standard antiderivative table doesn't directly apply. You need to rewrite using sin^2(x) = (1 - cos(2x))/2 or cos^2(x) = (1 + cos(2x))/2 before integrating. This turns a problem that looks impossible into one that takes thirty seconds. The same logic applies to products like sin(ax)cos(bx), which become sums of single trig terms after applying product-to-sum identities. Without those identities, you'd be stuck. There's also a limitation worth acknowledging straight up. The basic antiderivative formulas for trig functions only give you elementary results when the argument is simply x or a linear function of x. Once you have something like sin(x^2) or cos(e^x) in the integrand, you're no longer dealing with standard antiderivatives of trig functions — you're entering territory where no closed-form elementary antiderivative exists. Tools like Gradshteyn and Ryzhik or computer algebra systems can sometimes express these in terms of special functions like Fresnel integrals, but for most practical purposes, these integrals are left in their integral form or evaluated numerically. Don't waste time searching for an elementary antiderivative that isn't there. A common pitfall I see repeatedly involves definite integrals over intervals where the antiderivative has discontinuities. The function tan(x) has vertical asymptotes at x = /2 + n, and if your interval of integration crosses one of those points, the Fundamental Theorem of Calculus doesn't apply directly. You have to split the integral and evaluate limits approaching each asymptote from the appropriate side. I once saw a student plug in bounds across a discontinuity and get a numerically "correct" answer by accident, which reinforced the wrong intuition about why the method works.

Here's a specific example of the whole process working end to end. Consider the integral of 3sin(x) - 2cos(x) + sec^2(x) dx. You can break this into three separate integrals using linearity, then apply the standard formulas directly. The result is -3cos(x) - 2sin(x) + tan(x) + C. It seems almost too simple, and that's exactly the point — most problems at this level are this straightforward if you've internalized the basic pairs and can spot when a more complex integrand reduces to something elementary. When the integrand contains a composition like sin(3x) or cos(5x), you use the reverse chain rule, which in antidifferentiation terms means adjusting for the inner function's derivative. The integral of sin(3x) dx equals -1/3 cos(3x) + C because the derivative of 3x is 3, and you need to compensate by dividing by that factor. The same principle applies across all the basic pairs: whatever multiplier sits inside the argument requires a reciprocal adjustment outside the antiderivative. Substitution becomes necessary when the integrand includes a trig function multiplied by its own derivative, like sin(x)cos(x) dx, or when you have something like tan(x) dx that can be rewritten as sin(x)/cos(x) and solved with a u-substitution where u equals cos(x). Each of these cases follows a recognizable pattern once you've seen it a few times, and the goal is to build enough familiarity that you stop treating every problem as a fresh derivation and start recognizing which tool applies immediately.

The deeper you go into calculus, the more these basic antiderivatives surface in unexpected places — differential equations, Fourier analysis, probability distributions. Having them automated in your head so you're not constantly looking them up is worth the effort. Twenty minutes of deliberate practice on the core pairs will save you hours of frustration later when you're trying to focus on the actual problem instead of reconstructing foundational formulas from scratch.